The intersection between generative artificial intelligence and Web3 is one of the most active areas of research and development in crypto circles over the last few months. Decentralized compute, zero-knowledge AI, smaller foundation models, decentralized data networks, and AI-first chains are some of the recent trends that aim to enable Web3-native rails for AI workloads.

These trends are technological innovations that seek to bridge the worlds of Web3 and AI, representing a natural friction against the centralized nature of generative AI. While creating technological bridges with AI is foundational for the evolution of Web3, they don’t represent the only integration path for these technology trends.

What if the path for integrating Web3 and AI was financial instead of purely technical? It turns out that the programmable finance and capital formation capabilities of crypto could be useful for one of the biggest challenges facing the current generative AI market.

What challenge are we referring to? Nothing other than the funding challenges of open-source generative AI.

Open Source Generative AI Needs to Succeed

Despite the recent level of innovation in decentralized generative AI, the gap with centralized AI tech is increasing rather than decreasing. Many people agree that blockchains represent the best technology alternative to the increasing centralized AI control of large tech platforms. However, the adoption challenges for decentralized AI platforms are monumental.

Decentralized compute is a clear pillar for decentralized AI but proves impractical for pretraining and fine-tuning workloads that require GPUs in close proximity with access to datasets that often sit behind corporate firewalls. Zero-knowledge ML is too expensive to be practical in large foundation models and hasn’t seen any real demand in the market. Decentralized data marketplaces need to overcome the same issues that have prevented data marketplaces from becoming large tech businesses.

While decentralized AI strives to overcome these frictions, centralized alternatives are accelerating at a frantic pace, creating a scary gap between the two. The one trend that is keeping the hopes for a world in which decentralized AI can succeed is the rapid evolution of open-source generative AI.

All decentralized AI trends rely on a healthy open-source generative AI ecosystem, yet that ecosystem might not be as healthy as it seems.

Open Source Generative AI Has a Massive Funding Problem

In the last couple of years, we have witnessed an explosion of innovation in open-source large generative AI as an alternative to platforms such as OpenAI/Microsoft, Google or Anthropic. Meta has become a surprising undisputed champion of open-source generative AI with the release of the Llama models. Companies like Mistral have raised billions in venture funding, enterprise platforms like Databricks or Snowflake are pushing open-source models, and there is a growing number of open-source generative AI releases on a weekly basis.

While the momentum in open-source generative AI is strong, a more detailed analysis shows a different reality. Open-source generative AI is facing a massive funding issue. When it comes to large foundation models, only large companies such as Databricks, Snowflake, Meta or well-funded startups like Mistral are keeping up with the performance of large closed models. Most of the releases from other labs, like Databricks and Snowflake, are focused on optimized enterprise workloads, while most of the recent open-source research is focusing on complementary techniques rather than on new models.

The reason behind this phenomenon can be attributed to the astronomical costs of building large frontier models. Any pre-training cycle for a 20 billion-plus parameter model could cost between ten to a hundred million dollars and involves a multi-month process with many failed attempts. These costs fall outside the budget of most university labs. To make matters more interesting, many of the grants for AI university labs come from large tech incumbents, which then are the immediate beneficiaries of the outputs.

Making money with open source has historically been hard, and making money with open-source generative AI is hard at AI scale. As a result, open-source generative AI is experiencing a massive funding crunch that can create a serious gap with the AI incumbents.

Crypto capital for open-source generative AI

The capital formation primitives of crypto seem like one of the few viable alternatives to address the funding crunch in generative AI. Throughout its history, crypto tokens have been a primary vehicle for capital formation for Web3 projects through bull and bear market cycles. Could some of these principles be applied to open-source generative AI? There is certainly more than one interesting option.

1. Gitcoin Quadratic Funding
   • Gitcoin represents one of the most successful examples of funding open-source innovation in Web3. The quadratic funding mechanism pioneered by Gitcoin could apply directly to generative AI. Bringing native generative AI capabilities to Web3 is paramount for the evolution of the space, so it is natural to expect that generative AI projects will drive community attention.
   • Let’s say that a university AI lab needs to raise $10 million for pre-training an LLM based on novel architecture. Multiple DAOs and foundations can contribute to a Gitcoin grant that can also be matched by the grantors, creating a more efficient funding mechanism. This mechanism is far more efficient than the current alternatives in the market.
2. A New Open-Source Generative AI License
   • Funding open-source projects enables mechanisms in which the value created by those projects can benefit the original funding community. When it comes to Web3 and open generative AI, an interesting idea is to establish a license in which any commercial application using a model funded using Web3 tokens should contribute part of that revenue back in the form of that specific token. This mechanism can even be enforced via smart contracts.

Addressing a Systemic Risk to Open Generative AI

Financing vehicles for open-source AI are one of the most important challenges to address in the current generative AI landscape. Open source is traditionally hard to finance, and open-source generative AI is even more so, considering the expensive computational requirements.

Not enabling proper funding channels to foster open-source innovation in generative AI can create a systemic risk to the entire space as the balance will shift entirely to closed commercial platforms. Crypto has established some of the most sophisticated and battle-tested channels for funding open-source innovation. Maybe, the first bridge between Web3 and generative AI will be financial and not necessarily technical.

Source: CoinDesk